ECU professors, students develop AI tutor for fall classes
- ECU computer science professors and students created TheoTeach, an AI tutor designed to help students understand challenging, mathematical computer science concepts.
- Rather than simply providing answers, TheoTeach challenges students with questions, feedback and explanations, helping them build understanding and problem-solving skills.
- TheoTeach will be used in computer science classes this fall, with the team looking to expand its use to other programs, demonstrating how AI can be used as a tool to enhance education and student learning.
It’s not unheard of for students to get lost when complex mathematical and computational theories of computer science are discussed in class.
East Carolina University professors and students developed an AI-tutor system to better engage students and increase their knowledge of such topics. Dr. Moritz Dannhauer, assistant professor in the Department of Computer Science, plans to implement the system in his classes this fall.
“I want to get as many students engaged as possible,” Dannhauer said. “I teach very theoretical courses, so they’re very hard, surprisingly hard, and computer science students are usually very surprised how deep it goes. It’s very mathematical, and I’m trying to get it more digestible and deliver more in a more modern way.”
Dr. Moritz Dannhauer, left, assistant professor in the Department of Computer Science, and student Joshua Martinez work in class in the Austin Building. (Photo by Ken Buday)
TheoTeach
Dannhauer, computer science professor Dr. Nasseh Tabrizi and computer science students Joshua Martinez, Branlyn Roberson, Brice Hamlett and Forrest Fishel created TheoTeach. Dannhauer said it acts as a tutor, providing students with a secondary way to understand the lessons he’s teaching.
“Using TheoTeach to interact with the topic, it gives them a personal assistant basically that guides and challenges them to see if they truly understand the material,” Dannhauer said. “I am very proud that my students were able to build TheoTeach to help other students.”
From the outset, the team designed TheoTeach to be more than an artificial intelligence system that simply gives answers to questions the students pose, knowing that such an application doesn’t help students in the long run.
“AI will give them the answer, but they won’t learn anything,” Martinez said. “Students will use AI to replace their education, and replacing their education is the fastest way to lose their knowledge base.”
How it Works
The team developed TheoTeach based on Dannhauer’s lessons.
“What we constructed is a tutor that fills the gap between Dr. Dannhauer’s teaching and the students’ current state of understanding,” Martinez said.
If students get lost in the lesson, “then TheoTeach can be used to get them back on track, to remind them what the proper notation is, how to get to that proper notation through a different method, but still using the same material that Dr. Dannhauer is presenting, maybe from a slightly different angle,” Martinez said.
TheoTeach allows students to simulate computational machines such as Automata and Turing machines through other freely available online software and provides feedback on the designs so students can iteratively improve it, Dannhauer said.
As it provides explanations to student inqueries, the system prompts more engagement by asking the students topical follow-up questions. Dannhauer said that engagement supports student understanding of the lessons.
“It’s a helper that we can use in the classroom that really challenges the student. That’s the key. You can’t really learn if you’re not challenged,” he said.
What’s Next
Dannhauer and Martinez said initial testing with students showed promising results, allowing students to better understand classroom lectures.
“We noticed that there was a gap between what the students were getting and what Dr. Dannhauer wanted them to get, and hopefully this will be able to assist in that,” Martinez said.
Martinez received funding for the project through a College of Engineering and Technology mini grant as well as from the Department of Computer Science. The team presented its work during ECU’s Summer Undergraduate Research Symposium last month.
“When Joshua recently presented at the Summer Undergraduate Research Symposium, I saw the interest and feedback as this topic seems to be on the minds of educators and students alike,” Dannhauer said.
TheoTeach is expanding to other computer science courses. The team is looking to partner with other colleges and departments at ECU as well as other universities to deploy TheoTeach in classes beyond computer science.
Martinez said TheoTeach is an example of the positive side of artificial intelligence in education.
“It’s using AI as a tool,” he said. “And I think that’s the biggest thing.”